Umm, just for the record, I barely understand quats as it is - I can''t seem to find a simple explanation to gimme that "eurika" deal. BUT, I was just wondering if it is workable in 2 dimensions before I devote time to understanding it. I assume its a good concept to know, regardless, but I''m really interested in applying it to my game atm. I''ve been doing everything in my game with angles and its been annoying as nut - all the conversions and exceptions. So I was planning on doing a big convert to vectors and quats, if I understand it correctly, which should be a big improvement. Am I wrong or can you offer any advice/input? Thanks

You can use complex numbers for rotations in 2D, but I''m in ''study break'' at the moment so I can''t remember how.

quote:

---

Original post by neonfaktory
Umm, just for the record, I barely understand quats as it is - I can''t seem to find a simple explanation to gimme that "eurika" deal. BUT, I was just wondering if it is workable in 2 dimensions before I devote time to understanding it. I assume its a good concept to know, regardless, but I''m really interested in applying it to my game atm. I''ve been doing everything in my game with angles and its been annoying as nut - all the conversions and exceptions. So I was planning on doing a big convert to vectors and quats, if I understand it correctly, which should be a big improvement. Am I wrong or can you offer any advice/input? Thanks

---

Quats could work, you just never rotate along the axis that is held constant.

2 dimensional rotation has only 1 degree of freedom, and is an entirely different problem than 3d rotation. A scalar is sufficient for 2d rotation, so you probably won''t learn much by using quats for 2d rotation.

Well then that brings me to my next question. Is there some kind of all-encompassing angle/orinetation method I can use? Like I said, using actual "degrees" has been a huge pain. For example, when trying to orient to an object or orient based on movement direction, I have to do all these dumb little tests for if it is >90 facing away and getting the acos/asin/atan only giving you 0-180 degree range so you have to do special testing for which respective quadrant you are in and not to mention the whole issue of degree wrapping to make sure a constantly rotating, long living object does not exceed the range of the variable type holding the degree of rotation. GAH!

So my question, how do you guys handle this stuff?

Just have a direction vector that points in the direction you want to go. You can calculate this by (1,0) rotated by your angle.

Then, to see if something is facing something else, you can dot their direction vectors.